kill -9 again

> From: mreed@news.hcia.com (Michael T. Reed) > Subject: kill -9 again > > The **real** problem that we want to solve is the "end user" who > is dangerous enough with Unix to know that kill -9 will do the trick. > > Regardless of how often one preaches about not using kill -9, there > is always the occasional user (or programmer) that forgets or decides > to do it anyway, causing the Online engine to crash most of the time. Well, this sounds more like a system security issue. Being that the tbinit process is owned by 'informix', not just anybody should be able to send a kill -9 and have any affect. They won't have the authorization to pull it off. When we visit other installations, particulalry smaller ones, we see a lot of people logging in and doing work under the 'informix' login. I never understood why people do this and not just leave 'informix' alone and treat it the same way one treats the 'root' login. 'root' is all powerful at a system level, yet 'informix' is all powerful at the database level. On a side note regarding kill -9 (or more appropiately, tbmode -z), I don't know if I've ever seen a clean way to convince an sqlturbo that it is doing a Bad Thing. With 4GL 4.1, we were given the ability to pass an intr signal through the 4GL down into the engine, but sometimes you send the wrong thing off to sqlturbo and it seems unstoppable. Typically, these are really nasty queries that seem to feel like they have to join your 3 different million row tables using sequential scans on all of them. You know that this thing will never finish in your lifetime, you know it's not upadating anything save maybe an Informix generated internal temp table (which shouldn't be logged, but I don't know for sure), so there shouldn't be any major log activity. You also know that this one querie is sucking the life out of your system resources, yet no combinations of tbmodes, kills, bashes, or harsh words seem to stop the engine (save for the ever popular -9). All you can really do is shut down the engine to get these to stop, otherwise you sit back and watch the CPU time on sqlturbo soar to the triple digits. I'm not sure which the users like most... Has anyone else encountered Queries from Hell like this? Has anyone seem a tbmode WORK against a query like this? I don't think I've ever seen tbmode work against a query that wouldn't just as easily be stopped by killing the front end job. Will (villy@collinscomp.com)